Made available in DSpace on 2015-03-05T18:57:21Z (GMT). No. of bitstreams: 0 Previous issue date: 30 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Este trabalho investiga a função de reação de política monetária, seguindo a abordagem da Regra de Taylor para avaliar o desempenho dessa política, conduzida pela Reserva Federal (FED), pelo Banco do Japão (BOJ) e pelo Banco Central Europeu (ECB), durante o período selecionado para a pesquisa. Considerou-se para a análise, tanto para o FED como para o BOJ, o período de janeiro de 1990 até junho de 2008; enquanto que para o ECB, em virtude da constituição da Área do euro, a análise abrange janeiro de 1998 a junho de 2008. Inicialmente, é realizada a revisão da literatura sobre discricionariedade versus regras de política monetária, em que são apresentados alguns resultados empíricos sobre o uso de regras na condução da política monetária. Num segundo momento, faz-se uma abordagem sobre como estão estruturados os bancos centrais e os sistemas de pagamentos dos países selecionados. Além disso, traçam-se considerações sobre o regime monetário e cambial de cada economia e faz-se também uma breve retrospectiva da c / This work investigates the function of reaction of monetary policy following the approach of the Taylor Rule to evaluate the performance of this policy, lead for the Federal Reserve (FED), for the Bank of Japan (BOJ) and for European Central Bank (ECB), during the period selected for the research. It was considered for the analysis, as much for the FED how much for the BOJ, the period of January 1990 until June 2008; whereas for the ECB, in virtue of the constitution of the Euro Area, the analysis encloses January 1998 until June 2008. Initially, the revision of literature on discretion versus rules of monetary policy is made, where some empirical results on the use of rules in the conduction of the monetary policy are presented. At as a moment, approach becomes on as the central banks and the systems of payments of the selected countries are structuralized. Moreover, considerations are traced on the monetary and exchange regimen of each economy and become one brief retrospective of the management of the mone

Como a sustentabilidade da dívida é afetada por elevados e crescentes níveis de endividamento no Brasil? Esta dissertação introduz uma estrutura teórica e empírica para a estimação de limites endógenos para a dívida bruta do setor público, utilizando um modelo com credores neutros ao risco que avaliam a probabilidade de calote do governo em função do próprio nível de endividamento e da taxa de juros e um governo soberano que segue uma função de reação do superávit ao nível da dívida com propriedade de fadiga fiscal e choques exógenos. Nos dados de 2002 a 2018, as estimativas de limites para o endividamento bruto estão entre 130,8%, no cenário mais provável, até 151,4% do PIB, sob condições mais favoráveis. As taxas de juros estariam entre 22% e 38% a.a., garantindo a sustentabilidade fiscal do governo para os próximos anos, com ressalvas que 1) este é um máximo para a dívida, nada é discutido sobre um nível ótimo de endividamento; e 2) que a trajetória crescente da dívida em anos recentes requer uma atenção especial por parte das autoridades fiscais que desejarem se comprometer com sua sustentabilidade, ainda que ela se situe abaixo de seu máximo. / How the sustainability of public debt changes when rising indebtedness arises? This dissertation introduces the theoretical and empirical framework to estimate endogenous limits to the debt levels, that is, a model with risk-neutral lenders that account for the default probability as a function of indebtedness itself (thus the endogeneity) and the interest rates; and the fiscal authority, which follows a fiscal reaction function with the fiscal fatigue property and exogenous shocks. The estimates for the debt limits in the Brazilian economy stays around 130.8% and 151.4% of GDP, with real interest rates of 38% per year in the first case, using monthly data from 2002 to 2018. This ensures the fiscal sustainability for the next year, but it is worth noting that 1) this is the limit for debt levels; nothing is discussed about the optimal level for the economy as a whole; and 2) the path of rising indebtedness in the recent years needs to be strongly reviewed by whoever wishes to commit to fiscal sustainability in the next government.

The relations between monetary policies, asset prices, and economic growth are important and fundamental questions in macroeconomics. To address these issues, several empirical works have been conducted to investigate these relations. However, few of them have documented whether these relations differ across regimes. In this context, the general motivation of this thesis is to use dependent regime models to examine these relations for the Canadian case. Chapter one empirically analyzes the interest rate behaviour of the Canadian monetary authorities by taking into account the asymmetry in the loss function. We employ a switching regime framework using two estimation strategies: First, we follow Caner and Hansen (2004) Threshold approach. Under this procedure we estimate the threshold values, using the Taylor empirical rules. Second, we estimate the asymmetric policy reaction function following Favero and Rovelli’s (2003) approach. The results reveal that the monetary authorities showed asymmetric preferences and that its reaction function can be better modeled with a nonlinear model. The main contribution of this chapter is to successfully interpret the parameters associated with the Bank of Canada preferences, something that Rodriguez (2008) could not do. Chapter two tries to estimate the interest rate behaviour of the Canadian monetary authorities by expanding the arguments of the loss function for fluctuations in asset prices. Using the same methodology as in the first chapter, our findings suggest that the augmented nonlinear reaction function is a good fit for the data and gives new relevant insights into the influence of asset prices on Canadian monetary policy. These findings about the role of asset prices in the reaction function of the Bank of Canada provide relevant insights regarding the opportunities and limitations of incorporating financial indicators in monetary policy decision making. They also provide financial market participants, such as analysts, bankers and traders, with a better understanding of the impact of stock market index prices on Bank of Canada policy. Stock market stabilization plays a larger role in the interest rate decisions of the Bank of Canada than it is willing to admit. Chapter three provides new evidence on the relation between inflation, relative price variability and economic growth to a panel of Canadian provinces over the period 1981-2008. We use the Bick and Nautz (2008) modified version of Hansen’s (1999) Panel Threshold Model. The evidence strongly supports the view that the relationship between inflation and economic growth is nonlinear. Further investigation suggests that relative price variability is one of the important channels through which inflation affects economic performance in Canadian provinces. When taking into account the cross-section dependence, we find that the critical threshold value slightly changes. It is desirable to keep the inflation rate in a moderate inflation regime because it may be helpful for the achievement of sustainable economic growth. The results seem to indicate that inflation that is too high or too low may have detrimental effects on economic growth.

The relations between monetary policies, asset prices, and economic growth are important and fundamental questions in macroeconomics. To address these issues, several empirical works have been conducted to investigate these relations. However, few of them have documented whether these relations differ across regimes. In this context, the general motivation of this thesis is to use dependent regime models to examine these relations for the Canadian case. Chapter one empirically analyzes the interest rate behaviour of the Canadian monetary authorities by taking into account the asymmetry in the loss function. We employ a switching regime framework using two estimation strategies: First, we follow Caner and Hansen (2004) Threshold approach. Under this procedure we estimate the threshold values, using the Taylor empirical rules. Second, we estimate the asymmetric policy reaction function following Favero and Rovelli’s (2003) approach. The results reveal that the monetary authorities showed asymmetric preferences and that its reaction function can be better modeled with a nonlinear model. The main contribution of this chapter is to successfully interpret the parameters associated with the Bank of Canada preferences, something that Rodriguez (2008) could not do. Chapter two tries to estimate the interest rate behaviour of the Canadian monetary authorities by expanding the arguments of the loss function for fluctuations in asset prices. Using the same methodology as in the first chapter, our findings suggest that the augmented nonlinear reaction function is a good fit for the data and gives new relevant insights into the influence of asset prices on Canadian monetary policy. These findings about the role of asset prices in the reaction function of the Bank of Canada provide relevant insights regarding the opportunities and limitations of incorporating financial indicators in monetary policy decision making. They also provide financial market participants, such as analysts, bankers and traders, with a better understanding of the impact of stock market index prices on Bank of Canada policy. Stock market stabilization plays a larger role in the interest rate decisions of the Bank of Canada than it is willing to admit. Chapter three provides new evidence on the relation between inflation, relative price variability and economic growth to a panel of Canadian provinces over the period 1981-2008. We use the Bick and Nautz (2008) modified version of Hansen’s (1999) Panel Threshold Model. The evidence strongly supports the view that the relationship between inflation and economic growth is nonlinear. Further investigation suggests that relative price variability is one of the important channels through which inflation affects economic performance in Canadian provinces. When taking into account the cross-section dependence, we find that the critical threshold value slightly changes. It is desirable to keep the inflation rate in a moderate inflation regime because it may be helpful for the achievement of sustainable economic growth. The results seem to indicate that inflation that is too high or too low may have detrimental effects on economic growth.

Esta tese apresenta três ensaios sobre política monetária e fiscal dentro do atual regime de metas de inflação. O primeiro ensaio buscou estudar uma possível integração monetária-fiscal ao determinar uma regra ótima de política monetária com restrição fiscal, analisando os efeitos de diversas preferências sobre a regra ótima em função da alteração dos pesos dados para os desvios da razão superávit primário/PIB em relação à sua meta pré-estabelecida. Os resultados mostraram que a regra ótima obtida apresenta uma resposta negativa das taxas de juros aos choques na relação dívida/PIB. Ainda, superávits primários/PIB maiores permitiriam reduções maiores nas taxas de juros e proporcionais aos pesos que essa variávelobjetivo teria na função de perda social. Do ponto de vista tradicional do mecanismo de transmissão da política monetária, a resposta positiva das taxas de juros a uma desvalorização real do câmbio e a uma elevação do prêmio de risco seria mantida. Portanto, os resultados sugerem que a adoção de uma meta explícita para o superávit primário/PIB tem conseqüências positivas sobre a regra ótima de política monetária e para a redução da taxa de juros, bem como na eficiência do atual instrumento de política monetária. O segundo ensaio buscou analisar a relação risco default através do modelo de regressão beta, bem como os impactos que os superávits primários podem trazer sobre o prêmio de risco e, consequentemente, sobre o câmbio. Do ponto de vista da relação default risk, ancorada no modelo de Blanchard (2004/2005), as estimativas baseadas no modelo de regressão beta para as quatro relações propostas neste ensaio apresentaram sinais estatisticamente significativos e compatíveis com a teoria. O fato interessante nos resultados referente ao período do regime de metas de inflação é que as estimativas indicaram uma relação direta e forte entre o superávit primário/PIB e a probabilidade de default; evidências que destacam a importância dos efeitos indiretos que o superávit pode gerar sobre o juro doméstico. O terceiro ensaio analisou a dinâmica discreta da taxa de juros SELIC-meta definida nas reuniões do Comitê de Política Monetária (COPOM). Dois métodos foram utilizados para estudar a possibilidade de o COPOM reduzir/manter/aumentar a taxa de juros básica: probit binomial e probit multinomial. Os resultados mostraram que os desvios de inflação e o hiato do produto são variáveis relevantes para explicar as decisões do COPOM. O modelo probit binomial aplicado para os casos de aumento e redução da taxa SELIC-meta mostraram que a inclusão da variável fiscal gerou melhores resultados. Para o caso agregado, método probit multinomial, os resultados indicaram que a inclusão da variável fiscal combinada com a expectativa de inflação gerou os melhores resultados relativamente aos demais casos. Assim, a resposta do COPOM a resultados fiscais bem como às expectativas do mercado quanto à inflação demonstraram ser os sinais que devem ser observados pelo mercado. / This thesis presents three essays on monetary and fiscal policy of the current regimen of inflation targeting. The first essay searched to study an integration monetary-fiscal when determining an optimal rule of monetary policy with fiscal restriction, analyzing the effect of diverse preferences on the optimal rule in function of the alteration of the weights given for the deviations of the surplus primary as a fraction of GDP in relation to its established targets. The results show that the gotten optimal rule presents a negative reply of the interest rates to the shocks in the debtto- GDP ratio. Primary surplus still bigger would allow bigger reductions in the interest rates and proportional to the weights that this variable-objective would have in the function of social loss. Of the traditional point of view of the mechanism of transmission of the monetary policy, the positive reply of the interest rates to a real depreciation of the exchange and to a rise of the risk premium it would be kept. Therefore, the results suggest that the adoption of explicit targets for the primary surplus in percentage of the GDP has positive consequences on the optimal rule of monetary policy and for the reduction of the interest rates, as well as in the efficiency of the current instrument of monetary policy. The second essay searched to analyze the relation default risk through of the beta regression model, as well as the impacts that primary surplus can bring on the risk premium and, consequently, on the exchange rate. Of the point of view of the relation default risk, anchored in the model of Blanchard (2004/2005), the estimates based on the beta regression model for the four relations proposals in the study had presented significant and compatible signals with the theory. The interesting fact in the results referring to the period of the regimen of inflation targeting is that the estimates had indicated a negative and strong relation between the primary surplus/GDP and the probability of default, evidences that detaching the importance of the positive and indirect impact of the surplus in relation to the interests rate domestic. The third analyzes the discrete dynamics of the SELIC interest rates-target defined in the meetings of the Brazilian Monetary Policy Council (COPOM). Two methods were applied in order to study the possibility of COPOM to reduce/maintain/increase the interest rates: probit model and multinomial probit. It was verified that the deviations of inflation and the GDP gap must be considered importants variables to explain the COPOM’s decisions. The probit model was applied to the cases of the increases probabilies and reduces probabilities showing that the inclusion of a fiscal variable generates better results. To the aggregated case, multinominal probit method, the results indicates that the inclusion of a fiscal variables combined with the inflation expectations generates better results than other possibilities. So, the responses of COPOM to the fiscal results as well as inflation expectations were the reals signs to be considered for the market.

Esta tese apresenta três ensaios sobre política monetária e fiscal dentro do atual regime de metas de inflação. O primeiro ensaio buscou estudar uma possível integração monetária-fiscal ao determinar uma regra ótima de política monetária com restrição fiscal, analisando os efeitos de diversas preferências sobre a regra ótima em função da alteração dos pesos dados para os desvios da razão superávit primário/PIB em relação à sua meta pré-estabelecida. Os resultados mostraram que a regra ótima obtida apresenta uma resposta negativa das taxas de juros aos choques na relação dívida/PIB. Ainda, superávits primários/PIB maiores permitiriam reduções maiores nas taxas de juros e proporcionais aos pesos que essa variávelobjetivo teria na função de perda social. Do ponto de vista tradicional do mecanismo de transmissão da política monetária, a resposta positiva das taxas de juros a uma desvalorização real do câmbio e a uma elevação do prêmio de risco seria mantida. Portanto, os resultados sugerem que a adoção de uma meta explícita para o superávit primário/PIB tem conseqüências positivas sobre a regra ótima de política monetária e para a redução da taxa de juros, bem como na eficiência do atual instrumento de política monetária. O segundo ensaio buscou analisar a relação risco default através do modelo de regressão beta, bem como os impactos que os superávits primários podem trazer sobre o prêmio de risco e, consequentemente, sobre o câmbio. Do ponto de vista da relação default risk, ancorada no modelo de Blanchard (2004/2005), as estimativas baseadas no modelo de regressão beta para as quatro relações propostas neste ensaio apresentaram sinais estatisticamente significativos e compatíveis com a teoria. O fato interessante nos resultados referente ao período do regime de metas de inflação é que as estimativas indicaram uma relação direta e forte entre o superávit primário/PIB e a probabilidade de default; evidências que destacam a importância dos efeitos indiretos que o superávit pode gerar sobre o juro doméstico. O terceiro ensaio analisou a dinâmica discreta da taxa de juros SELIC-meta definida nas reuniões do Comitê de Política Monetária (COPOM). Dois métodos foram utilizados para estudar a possibilidade de o COPOM reduzir/manter/aumentar a taxa de juros básica: probit binomial e probit multinomial. Os resultados mostraram que os desvios de inflação e o hiato do produto são variáveis relevantes para explicar as decisões do COPOM. O modelo probit binomial aplicado para os casos de aumento e redução da taxa SELIC-meta mostraram que a inclusão da variável fiscal gerou melhores resultados. Para o caso agregado, método probit multinomial, os resultados indicaram que a inclusão da variável fiscal combinada com a expectativa de inflação gerou os melhores resultados relativamente aos demais casos. Assim, a resposta do COPOM a resultados fiscais bem como às expectativas do mercado quanto à inflação demonstraram ser os sinais que devem ser observados pelo mercado. / This thesis presents three essays on monetary and fiscal policy of the current regimen of inflation targeting. The first essay searched to study an integration monetary-fiscal when determining an optimal rule of monetary policy with fiscal restriction, analyzing the effect of diverse preferences on the optimal rule in function of the alteration of the weights given for the deviations of the surplus primary as a fraction of GDP in relation to its established targets. The results show that the gotten optimal rule presents a negative reply of the interest rates to the shocks in the debtto- GDP ratio. Primary surplus still bigger would allow bigger reductions in the interest rates and proportional to the weights that this variable-objective would have in the function of social loss. Of the traditional point of view of the mechanism of transmission of the monetary policy, the positive reply of the interest rates to a real depreciation of the exchange and to a rise of the risk premium it would be kept. Therefore, the results suggest that the adoption of explicit targets for the primary surplus in percentage of the GDP has positive consequences on the optimal rule of monetary policy and for the reduction of the interest rates, as well as in the efficiency of the current instrument of monetary policy. The second essay searched to analyze the relation default risk through of the beta regression model, as well as the impacts that primary surplus can bring on the risk premium and, consequently, on the exchange rate. Of the point of view of the relation default risk, anchored in the model of Blanchard (2004/2005), the estimates based on the beta regression model for the four relations proposals in the study had presented significant and compatible signals with the theory. The interesting fact in the results referring to the period of the regimen of inflation targeting is that the estimates had indicated a negative and strong relation between the primary surplus/GDP and the probability of default, evidences that detaching the importance of the positive and indirect impact of the surplus in relation to the interests rate domestic. The third analyzes the discrete dynamics of the SELIC interest rates-target defined in the meetings of the Brazilian Monetary Policy Council (COPOM). Two methods were applied in order to study the possibility of COPOM to reduce/maintain/increase the interest rates: probit model and multinomial probit. It was verified that the deviations of inflation and the GDP gap must be considered importants variables to explain the COPOM’s decisions. The probit model was applied to the cases of the increases probabilies and reduces probabilities showing that the inclusion of a fiscal variable generates better results. To the aggregated case, multinominal probit method, the results indicates that the inclusion of a fiscal variables combined with the inflation expectations generates better results than other possibilities. So, the responses of COPOM to the fiscal results as well as inflation expectations were the reals signs to be considered for the market.

Esta tese apresenta três ensaios sobre política monetária e fiscal dentro do atual regime de metas de inflação. O primeiro ensaio buscou estudar uma possível integração monetária-fiscal ao determinar uma regra ótima de política monetária com restrição fiscal, analisando os efeitos de diversas preferências sobre a regra ótima em função da alteração dos pesos dados para os desvios da razão superávit primário/PIB em relação à sua meta pré-estabelecida. Os resultados mostraram que a regra ótima obtida apresenta uma resposta negativa das taxas de juros aos choques na relação dívida/PIB. Ainda, superávits primários/PIB maiores permitiriam reduções maiores nas taxas de juros e proporcionais aos pesos que essa variávelobjetivo teria na função de perda social. Do ponto de vista tradicional do mecanismo de transmissão da política monetária, a resposta positiva das taxas de juros a uma desvalorização real do câmbio e a uma elevação do prêmio de risco seria mantida. Portanto, os resultados sugerem que a adoção de uma meta explícita para o superávit primário/PIB tem conseqüências positivas sobre a regra ótima de política monetária e para a redução da taxa de juros, bem como na eficiência do atual instrumento de política monetária. O segundo ensaio buscou analisar a relação risco default através do modelo de regressão beta, bem como os impactos que os superávits primários podem trazer sobre o prêmio de risco e, consequentemente, sobre o câmbio. Do ponto de vista da relação default risk, ancorada no modelo de Blanchard (2004/2005), as estimativas baseadas no modelo de regressão beta para as quatro relações propostas neste ensaio apresentaram sinais estatisticamente significativos e compatíveis com a teoria. O fato interessante nos resultados referente ao período do regime de metas de inflação é que as estimativas indicaram uma relação direta e forte entre o superávit primário/PIB e a probabilidade de default; evidências que destacam a importância dos efeitos indiretos que o superávit pode gerar sobre o juro doméstico. O terceiro ensaio analisou a dinâmica discreta da taxa de juros SELIC-meta definida nas reuniões do Comitê de Política Monetária (COPOM). Dois métodos foram utilizados para estudar a possibilidade de o COPOM reduzir/manter/aumentar a taxa de juros básica: probit binomial e probit multinomial. Os resultados mostraram que os desvios de inflação e o hiato do produto são variáveis relevantes para explicar as decisões do COPOM. O modelo probit binomial aplicado para os casos de aumento e redução da taxa SELIC-meta mostraram que a inclusão da variável fiscal gerou melhores resultados. Para o caso agregado, método probit multinomial, os resultados indicaram que a inclusão da variável fiscal combinada com a expectativa de inflação gerou os melhores resultados relativamente aos demais casos. Assim, a resposta do COPOM a resultados fiscais bem como às expectativas do mercado quanto à inflação demonstraram ser os sinais que devem ser observados pelo mercado. / This thesis presents three essays on monetary and fiscal policy of the current regimen of inflation targeting. The first essay searched to study an integration monetary-fiscal when determining an optimal rule of monetary policy with fiscal restriction, analyzing the effect of diverse preferences on the optimal rule in function of the alteration of the weights given for the deviations of the surplus primary as a fraction of GDP in relation to its established targets. The results show that the gotten optimal rule presents a negative reply of the interest rates to the shocks in the debtto- GDP ratio. Primary surplus still bigger would allow bigger reductions in the interest rates and proportional to the weights that this variable-objective would have in the function of social loss. Of the traditional point of view of the mechanism of transmission of the monetary policy, the positive reply of the interest rates to a real depreciation of the exchange and to a rise of the risk premium it would be kept. Therefore, the results suggest that the adoption of explicit targets for the primary surplus in percentage of the GDP has positive consequences on the optimal rule of monetary policy and for the reduction of the interest rates, as well as in the efficiency of the current instrument of monetary policy. The second essay searched to analyze the relation default risk through of the beta regression model, as well as the impacts that primary surplus can bring on the risk premium and, consequently, on the exchange rate. Of the point of view of the relation default risk, anchored in the model of Blanchard (2004/2005), the estimates based on the beta regression model for the four relations proposals in the study had presented significant and compatible signals with the theory. The interesting fact in the results referring to the period of the regimen of inflation targeting is that the estimates had indicated a negative and strong relation between the primary surplus/GDP and the probability of default, evidences that detaching the importance of the positive and indirect impact of the surplus in relation to the interests rate domestic. The third analyzes the discrete dynamics of the SELIC interest rates-target defined in the meetings of the Brazilian Monetary Policy Council (COPOM). Two methods were applied in order to study the possibility of COPOM to reduce/maintain/increase the interest rates: probit model and multinomial probit. It was verified that the deviations of inflation and the GDP gap must be considered importants variables to explain the COPOM’s decisions. The probit model was applied to the cases of the increases probabilies and reduces probabilities showing that the inclusion of a fiscal variable generates better results. To the aggregated case, multinominal probit method, the results indicates that the inclusion of a fiscal variables combined with the inflation expectations generates better results than other possibilities. So, the responses of COPOM to the fiscal results as well as inflation expectations were the reals signs to be considered for the market.

Como a sustentabilidade da dívida é afetada por elevados e crescentes níveis de endividamento no Brasil? Esta dissertação introduz uma estrutura teórica e empírica para a estimação de limites endógenos para a dívida bruta do setor público, utilizando um modelo com credores neutros ao risco que avaliam a probabilidade de calote do governo em função do próprio nível de endividamento e da taxa de juros e um governo soberano que segue uma função de reação do superávit ao nível da dívida com propriedade de fadiga fiscal e choques exógenos. Nos dados de 2002 a 2018, as estimativas de limites para o endividamento bruto estão entre 130,8%, no cenário mais provável, até 151,4% do PIB, sob condições mais favoráveis. As taxas de juros estariam entre 22% e 38% a.a., garantindo a sustentabilidade fiscal do governo para os próximos anos, com ressalvas que 1) este é um máximo para a dívida, nada é discutido sobre um nível ótimo de endividamento; e 2) que a trajetória crescente da dívida em anos recentes requer uma atenção especial por parte das autoridades fiscais que desejarem se comprometer com sua sustentabilidade, ainda que ela se situe abaixo de seu máximo. / How the sustainability of public debt changes when rising indebtedness arises? This dissertation introduces the theoretical and empirical framework to estimate endogenous limits to the debt levels, that is, a model with risk-neutral lenders that account for the default probability as a function of indebtedness itself (thus the endogeneity) and the interest rates; and the fiscal authority, which follows a fiscal reaction function with the fiscal fatigue property and exogenous shocks. The estimates for the debt limits in the Brazilian economy stays around 130.8% and 151.4% of GDP, with real interest rates of 38% per year in the first case, using monthly data from 2002 to 2018. This ensures the fiscal sustainability for the next year, but it is worth noting that 1) this is the limit for debt levels; nothing is discussed about the optimal level for the economy as a whole; and 2) the path of rising indebtedness in the recent years needs to be strongly reviewed by whoever wishes to commit to fiscal sustainability in the next government.

The relations between monetary policies, asset prices, and economic growth are important and fundamental questions in macroeconomics. To address these issues, several empirical works have been conducted to investigate these relations. However, few of them have documented whether these relations differ across regimes. In this context, the general motivation of this thesis is to use dependent regime models to examine these relations for the Canadian case. Chapter one empirically analyzes the interest rate behaviour of the Canadian monetary authorities by taking into account the asymmetry in the loss function. We employ a switching regime framework using two estimation strategies: First, we follow Caner and Hansen (2004) Threshold approach. Under this procedure we estimate the threshold values, using the Taylor empirical rules. Second, we estimate the asymmetric policy reaction function following Favero and Rovelli’s (2003) approach. The results reveal that the monetary authorities showed asymmetric preferences and that its reaction function can be better modeled with a nonlinear model. The main contribution of this chapter is to successfully interpret the parameters associated with the Bank of Canada preferences, something that Rodriguez (2008) could not do. Chapter two tries to estimate the interest rate behaviour of the Canadian monetary authorities by expanding the arguments of the loss function for fluctuations in asset prices. Using the same methodology as in the first chapter, our findings suggest that the augmented nonlinear reaction function is a good fit for the data and gives new relevant insights into the influence of asset prices on Canadian monetary policy. These findings about the role of asset prices in the reaction function of the Bank of Canada provide relevant insights regarding the opportunities and limitations of incorporating financial indicators in monetary policy decision making. They also provide financial market participants, such as analysts, bankers and traders, with a better understanding of the impact of stock market index prices on Bank of Canada policy. Stock market stabilization plays a larger role in the interest rate decisions of the Bank of Canada than it is willing to admit. Chapter three provides new evidence on the relation between inflation, relative price variability and economic growth to a panel of Canadian provinces over the period 1981-2008. We use the Bick and Nautz (2008) modified version of Hansen’s (1999) Panel Threshold Model. The evidence strongly supports the view that the relationship between inflation and economic growth is nonlinear. Further investigation suggests that relative price variability is one of the important channels through which inflation affects economic performance in Canadian provinces. When taking into account the cross-section dependence, we find that the critical threshold value slightly changes. It is desirable to keep the inflation rate in a moderate inflation regime because it may be helpful for the achievement of sustainable economic growth. The results seem to indicate that inflation that is too high or too low may have detrimental effects on economic growth.

Les Banques centrales occupent une position essentielle au coeur des économies contemporaines et font l'objet de débats animés sur ce que devraient être le rôle, les objectifs et le cadre institutionnel d'une Banque centrale. Face à une multitude d'analyses normatives concernées par ce qui doit être, on note une absence relative d'analyses positives. L'objectif de cette thèse est de contribuer à éclairer la réflexion sur les choix de politique monétaire mis en oeuvre par les Banques centrales. Le cadre d'application empirique retenu pour notre recherche est celui de l'Union économique et monétaire (UEM). La place importante de cet ensemble économique sur la scène internationale et son cadre institutionnel particulier suffisent à justifier l'intérêt d'une étude du comportement de la Banque Centrale Européenne (BCE). Les nombreux débats qui ont accompagné la mise en place de cette institution encore relativement jeune et ceux qui se poursuivent aujourd'hui renforcent cet intérêt. Dans la littérature économique, deux lignes de conduite ont principalement été retenues pour analyser les choix effectifs de la politique monétaire des banques centrales. La première approche, que nous avons qualifiée d' « économique », consiste en la formulation et l'estimation des fonctions de réaction de politique monétaire déduites d'un comportement d'optimisation ne prenant en compte que des variables économiques. Cette approche est fondée sur une vision globalisante de l'autorité monétaire... / Central banks occupy a key position at the heart of contemporary economies and are the subject of lively debates about what should be the role, the objectives and the institutional framework of a central bank. In the presence of a multitude of normative analyses concerned with what should be, there is a relative lack of positive tests. The objective of this thesis is to contribute to elucidate the reflections on monetary policy choices implemented by central banks. The empirical application framework used for this research is that of the European Monetary Union (EMU). The crucial position of this economic entity on the international scene and its particular institutional framework suffice to justify the interest of analysing the behaviour of the European Central Bank (ECB). The numerous debates that accompanied the establishment of this still relatively young institution and those who continue today reinforce this interest. In economic literature, two guidelines were mainly used to analyse the actual choices of the monetary policy of central banks. The first approach, which we call "economic", is the formulation and estimation of reaction functions of monetary policy derived from optimising behaviour taking into account economic variables only. This approach relies on a globalising vision of the monetary authority...